67 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Journal of Health Disparities Research and Practice Volume 10, Issue 4, Winter 2017, pp 67-76 2011 Center for Health Disparities Research School of Community Health Sciences University of Nevada, Las Vegas Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Jake R. Morgan, Boston University Mari-Lynn Drainoni, Boston University Cindy Christiansen, Boston University Tamar F. Barlam, Boston University **Corresponding Author: Jake Morgan, Email: jakem@bu.edu ABSTRACT Objective: To examine whether patient and provider concordance by sex or race predicts antibiotic prescribing for acute bronchitis. Study setting: General Internal Medicine and Family Medicine adult clinics at a large safety-net hospital. Study design: We used a logistic model of prescribing as a function of race and sex concordance. Data extraction: Data were extracted from de-identified patient records for those with an acute bronchitis visit between 2008 and 2010. Principal findings: 71% (95% CI 68%-73%) of visits resulted in an antibiotic prescription. Patients in race-concordant visits were 17% (95% CI 8%-25%) less likely to receive a prescription. Conclusions: Race-concordant outpatient visits were associated with more appropriate antibiotic use. Examination of characteristics such as concordance can improve our understanding of the prescribing process and inform stewardship efforts. Keywords: Concordance, Antibiotics, Over-prescribing INTRODUCTION Antibiotics are prescribed for acute bronchitis 60%-80% of the time despite clinical guidelines strongly advocating against the practice (Centers for Disease Control and Prevention 2014). Efforts to combat overprescribing in outpatient settings have proven largely unsuccessful (Davey et al. 2013). This lack of success is worrisome given increasing bacterial resistance; the CDC estimated that over 2 million individuals were infected with antibiotic resistant bacteria in 2013 (Centers for Disease Control and Prevention 2014). Better characterizing predictors of the prescribing decision may inform the development of more effective antibiotic stewardship interventions. Previous research has demonstrated associations between prescribing and patient factors including age, sex, and comorbidities (Barlam et al. 2015; Kroening-Roche et al. 2012);
68 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for physician factors such as practice years, teaching experience, practice volume, and specialty (Barlam et al. 2015; Steinman, Landefeld, and Gonzales 2003); and practice variation across geographic regions (Barlam et al. 2015). The patient-provider relationship is potentially an important determinant of antibiotic prescribing decisions. Patient expectations (Coenen et al., 2013), the provider s perception of a patient s preferences (Tonkin-Crine et al., 2014), and communication (Mangione-Smith, Stivers, Elliott, McDonald, & Heritage, 2003) have been suggested as catalysts for antibiotic overuse. Race and sex may impact the patient-provider relationship by affecting communication during the clinical encounter. Providers knowledgeable about a patient s culture may be better able to clearly explain medical diagnoses, and a patient with a sense of shared experience may be more likely to trust their provider s judgment (Cooper-Patrick et al. 1999). Studies of cross-cultural communication have found, for example, that black patients treated by black providers rated the visits as more participatory than when treated by white providers (Cooper-Patrick et al. 1999). Similar trends have been observed in studies of gender concordance where care of a female patient by a female provider is more likely to be patient centered as measured by the Davis Observation Code (Bertakis and Azari 2012). However, other research has failed to find significant associations between medical care and patient-provider race or sex concordance (Jerant et al. 2011). Measures of sex and race concordance have not yet been used to examine inappropriate antibiotic prescribing for acute bronchitis. The objective of this study was to examine the effect of race and sex concordance with other predictors on prescribing in cases of acute bronchitis. METHODS Design We used retrospective electronic medical record (EMR) data to identify outpatient visits for acute bronchitis at a large, urban safety-net hospital. Visit data on patients and providers were abstracted from the EMR via the hospital s clinical data warehouse. Provider characteristics not in the medical record were abstracted from publicly available physician profiles. Study staff reviewed the professional headshot, educational history, and provider name to estimate race. The Boston University Institutional Review Board approved all study activities. Visits Visits for acute bronchitis (ICD-9-CM codes 466.x, 490) conducted between 2008 and 2010 in the outpatient clinic by either Family Medicine or General Internal Medicine attending physicians were included if the clinician had at least 10 visits for respiratory tract infections in the dataset. All visits included in the study involved patients over 18 years of age. Variables The primary outcome was the prescription of an antibiotic for an acute bronchitis visit. We reviewed a sample of charts to confirm that the ICD-9 diagnosis of bronchitis matched the documented clinical indicators. Visits where patients were diagnosed with both acute bronchitis and another infection that might warrant an antibiotic were excluded because it was unclear which diagnosis was considered in the prescribing decision. In general, antibiotics are not indicated for acute bronchitis (Centers for Disease Control and Prevention 2014). The primary independent variables were: 1) race concordance between patient and provider and 2) sex concordance between patient and provider. Race was recorded in six categories white, black, Hispanic/Latino, Asian, Middle Eastern or other/unknown. Patient race
69 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for was based on self-report. The EMR combines race and ethnicity, thus Hispanic/Latino was included in this category. To determine provider race, a five-person panel used a consensus method to assign one of the six racial categories based on a photo (Ellis and Deregowski 1981), surname (Elliott et al. 2008), and/or medical school (Elliott et al. 2008). This method was informed by previous research demonstrating over 95% agreement between observer-recorded race and self-reported race (Sohn et al., 2006). Panel members were given a card for each race category to hold up each time they were presented with a photo, surname, and school. Members were not prompted beforehand, and used their own judgement to assign race based only on photo, surname, and school in order to solicit their most natural response. If there was disagreement, the panel was given the opportunity to reach consensus. When consensus could not be reached, race was assigned as other/unknown. Patients or providers with a race of other/unknown were excluded from the analysis as concordance could not be established. Provider sex was recorded as male or female. Other chosen covariates were those previously found to be associated with antibiotic prescribing in the literature and potentially confound the relationship between concordance and prescribing, including patient factors age (Barlam, Soria-Saucedo, Cabral, & Kazis, 2016), sex (Barlam, Morgan, Wetzler, Christiansen, & Drainoni, 2015), insurance status (Vaz et al., 2015), race (Hicks et al., 2015), and history of chronic lung or heart disease (Shapiro, Hicks, Pavia, & Hersh, 2013) and provider factors age (Tell, Engström, & Mölstad, 2015), sex (Jerant, Bertakis, Fenton, Tancredi, & Franks, 2011), race (Cooper-Patrick et al., 1999), specialty (Family vs. General Internal Medicine) (Steinman, Landefeld, & Gonzales, 2003), and U.S. region of residency training (Barlam et al., 2015). Race and sex covariates reflect only the race and sex of the individual patient or provider and are different than the concordance variable, which measures the agreement in either sex or race between patients and providers. Statistical Analysis We performed chi-square tests to analyze differences between patient and provider characteristics in visits resulting in an antibiotic prescription compared with visits where no antibiotic was prescribed. Multivariable regression was used to examine the predictors of antibiotic prescribing, and we conducted an effect modification analysis to evaluate the effect of the presence of both race and sex concordance compared to the independent effect of each. Relative risks from the multivariable models were estimated with a binomial complementary loglog (cloglog) model including random effects to control for individual provider heterogeneity and to ensure that any findings around concordance were not artifacts of individual prescribing practices. Methods of estimating the relative risk from the cloglog model are described elsewhere (Cummings 2009). We conducted several sensitivity analyses to evaluate model fit. Because our data did not include potentially relevant clinical variables, such as underlying immunosuppression, we evaluated adding a measure of patient complexity, the Charlson Comorbidity Index (CCI) (Charlson et al. 1987; Grant et al. 2011). The CCI was initially developed to classify comorbid conditions which may affect mortality, and has since been more widely used to control for comorbidities and patient complexity more broadly (Charlson et al. 1987; Grant et al. 2011). We evaluated the increased chance (Sohn et al., 2006) of misclassifying Hispanic/Latino patients and providers by conducting a sensitivity analysis excluding these individuals and comparing the results to the full sample, and also evaluated inclusion of fixed effects for year. Sensitivity
70 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for analyses using other regression link functions and using the model with a random subsample of 1/3 of the dataset further validated the results. All analyses were conducted in Stata 13.1. RESULTS There were 1,177 visits for acute bronchitis included in our study. These visits were overseen by 70 unique providers. Two providers and 81 patients had a race of other/unknown and were excluded from the analysis. There were no other missing data. Most visits were with patients who were female, non-white, and publicly insured through Medicare or Medicaid (Table 1). The majority of physicians conducting the visits were female and white. Overall, 831 (71%; 95% CI 68%-73%) visits resulted in an antibiotic prescription. Visits in which patients received an antibiotic were associated with patients of younger age (72% vs. 64% under age 65), female sex (75% vs. 63%), and a history of lung disease (75% vs. 67%) (Table 1). An antibiotic was more often prescribed when the visit was conducted by a physician who was female (76% vs. 65% for males), under age 45 (78% vs. 68% for 45 or older), and specializing in Family Medicine (84% vs. 67% for General Internal Medicine). In the primary adjusted model, visits where the patient and provider shared the same race had an 17% lower risk of resulting in an antibiotic prescription compared to visits which were not race concordant (RR=0.83, 95% CI 0.75-0.92) (Table 2). This result was unchanged if Hispanic/Latino patients were excluded. Sex-concordant visits were not associated with antibiotic prescriptions (RR=0.94, 95% CI 0.88-1.01) while visits with white patients (RR=1.13, 95% CI 1.06-1.21) resulted in more prescriptions. Visits with male patients were less likely to be prescribed (RR=0.89, 95% CI 0.82-0.97) and those with a history of heart disease (RR=1.13, 95% CI 1.01-1.26) were more likely. There was a higher risk of prescribing (RR=1.29, 95% CI 1.10-1.50) if the visit was conducted by a provider who had completed residency in the West, although this result was based on a small sample. The effect modification analysis included a term interacting race and sex concordance (Table 2). After controlling for other confounders, the interaction between race and sex concordance did not significantly predict an antibiotic prescription (RR=0.93, 95% CI 0.77-1.12). As in the primary analysis, sex-concordant visits were not associated with antibiotic prescribing (RR=0.97, 95% CI 0.87-1.07). The effect modification analysis did not change the size, direction, or significance of any of the estimated effects of confounders. Our results did not substantially change in sensitivity analyses. Inclusion of the CCI in order to measure patient complexity did not change the results and was insignificant (p=0.3) and was thus excluded in favor of highlighting the effect of comorbidities often cited as related to antibiotic prescription, chronic lung disease and heart disease. Year fixed effects were insignificant and did not affect the overall results, suggesting prescribing trends were stable over our observation period.
71 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Table 1: Prescribing at acute bronchitis visits Antibiotic Prescription Yes No p-value Total n (%) n (%) Race Concordant Visit 454 287 (63.2) 167 (36.8) Race Non-Concordant Visit 723 544 (75.2) 179 (24.8) Sex Concordant Visit 327 221 (67.6) 106 (32.4) Sex Non-Concordant Visit 850 610 (71.8) 240 (28.2) 0.16 Visit-Level Patient Characteristics Sex Male 447 281 (62.9) 166 (37.1) Female 730 550 (75.3) 180 (24.7) Age group Under 65 years old 945 682 (72.2) 263 (27.8) Over 65 years old 232 149 (64.0) 83 (36.0) 0.02 Race White 551 382 (69.3) 169 (30.7) Black 413 298 (72.2) 115 (27.9) Asian 33 21 (63.6) 12 (36.4) 0.52 Middle Eastern 8 5 (62.5) 3 (37.5) Hispanic/Latino 170 124 (72.9) 46 (27.1) Insurance Commercial 490 335 (68.4) 155 (31.6) Medicare 306 220 (71.9) 86 (28.1) Medicaid 277 202 (72.9) 75 (27.1) 0.49 Other Insurance 54 41 (75.9) 13 (24.1) Uninsured 50 33 (66.0) 17 (34.0) History of Heart Disease 109 82 (75.2) 27 (24.8) 0.27 History of Chronic Lung Disease 550 411 (74.7) 139 (25.3) Visit Level Provider Characteristics Sex Male 587 382 (65.0) 205 (35.0) Female 590 449 (76.1) 141 (23.9) Age group Under 45 years old 467 366 (78.4) 101 (21.6) Over 45 years old 710 465 (65.5) 245 (34.5) Race White 821 557 (67.8) 264 (32.2) Non-White 356 274 (77.0) 82 (23.0) Provider Residency Training Northeast 1,047 741 (70.8) 306 (29.2) South 38 35 (92.1) 3 (7.9) Midwest 76 40 (52.6) 36 (47.4) West 16 15 (93.8) 1 (6.3) Specialty of provider Family Medicine 267 223 (83.5) 44 (16.5) General Internal Medicine 910 608 (66.8) 302 (33.2)
72 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Table 2: Risk ratios of receiving an antibiotic prescription Primary analysis Effect modification analysis Risk Ratio (95% CI) Risk Ratio (95% CI) Race and sex concordant visit - - 0.93 (0.77-1.12) Race concordant visit 0.83 (0.75-0.92) 0.87 (0.76-0.99) Sex concordant visit 0.94 (0.88-1.01) 0.97 (0.87-1.07) Visit-Level Patient Characteristics Sex Male 0.89 (0.82-0.97) 0.89 (0.82-0.97) Female 1.00 (Reference) 1.00 (Reference) Age group Under 65 years old 1.00 (Reference) 1.00 (Reference) Over 65 years old 0.84 (0.70-1.02) 0.84 (0.69-1.02) Race* White 1.13 (1.06-1.21) 1.13 (1.06-1.21) Black 1.00 (Reference) 1.00 (Reference) Insurance Commercial 1.00 (Reference) 1.00 (Reference) Medicare 1.06 (0.94-1.20) 1.06 (0.94-1.20) Medicaid 1.00 (0.90-1.10) 1.00 (0.91-1.11) Other insurance 1.07 (0.93-1.22) 1.07 (0.94-1.23) Uninsured 0.98 (0.87-1.11) 0.99 (0.87-1.12) History of heart disease 1.13 (1.01-1.26) 1.13 (1.01-1.26) History of chronic lung disease 1.08 (0.98-1.18) 1.07 (0.98-1.18) Visit Level Provider Characteristics Sex Male 0.92 (0.76-1.10) 0.91 (0.76-1.09) Female 1.00 (Reference) 1.00 (Reference) Age group Under 45 years old 1.00 (Reference) 1.00 (Reference) Over 45 years old 0.95 (0.80-1.12) 0.95 (0.80-1.13) Race White 1.02 (0.81-1.28) 1.02 (0.81-1.28) Non-white 1.00 (Reference) 1.00 (Reference) Provider residency training Northeast 1.00 (Reference) 1.00 (Reference) South 1.24 (0.99-1.57) 1.24 (0.99-1.57) Midwest 0.90 (0.55-1.47) 0.91 (0.55-1.48) West 1.29 (1.10-1.50) 1.28 (1.09-1.51) Specialty of provider Family medicine 1.16 (0.97-1.38) 1.16 (0.97-1.38) General internal medicine 1.00 (Reference) 1.00 (Reference) *Race categories were collapsed to white vs. non-white due to sample size. DISCUSSION Our study demonstrated that race concordance played a role in appropriate antibiotic prescribing. We found that for acute bronchitis visits where the provider and patient were of the same race were 17% less likely to result in an antibiotic prescription. We did not find a significant relationship for sex concordance. The effect modification analysis yielded similar results, with only race concordance (and not sex or dual race and sex concordance) predicting antibiotic prescribing. This suggests that any effect of a combined race/sex concordance variable
73 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for is driven by the association between race concordance and prescribing. Previous studies have focused on the relationship between either provider factors or patient factors and inappropriate prescribing (Kroening-Roche et al. 2012). To our knowledge, sex and race-concordance have never been used in an explanatory model of antibiotic prescribing, and the results of this study suggest that further concordance research would be fruitful. Our findings can be used to inform efforts to improve antimicrobial stewardship interventions, which to date have had limited success. Currently, interventions have focused on provider knowledge and behavior (Davey et al. 2013). Interventions directed towards the patient have been less frequent and often are limited to provision of informational materials (Davey et al. 2013). Our findings indicate that the patient-provider relationship may be a useful area to explore possible interventions, whether through provider training aimed at improving crosscultural communication, relationship-building, or patient-centered communication which engages and educates the patient. Several systematic reviews have suggested that patient-provider communication may be an important area of focus, but there is no consensus on the area of communication that is weakest, or the best way to improve it (Davey et al. 2013). Our results suggest that cross-cultural communication can affect prescribing. This is particularly important in urban safety-net medical centers, which often serve racially diverse and socioeconomically disadvantaged populations who may experience disparities in care (Fiscella and Williams 2004). Our analysis is not able to determine whether the prescription of antibiotics, or the effect of race-concordance, is primarily a patient- or provider-driven phenomenon. It could be that a provider was less able to convince a non-concordant patient that antibiotics were unnecessary because he or she could not culturally connect with the patient. Or perhaps patients are more likely to trust, and take advice from, providers with whom they feel a sense of shared identity. Future research should consider employing strategies identified in studies on health care disparity, including cross-cultural communication and awareness, to better understand the dynamics of the patient-provider relationship in prescribing. Our study has several limitations. This study did not collect provider race data directly from physicians. It is possible several providers were misclassified, affecting our results. However, previous research has demonstrated strong agreement between observed and selfreported race (Sohn et al. 2006). Hispanic/Latino origin is an important cultural variable, but is difficult to assess with certainty via clinical records or visual inspection. It is possible that some Hispanic/Latino providers may have been misclassified as non-hispanic/latino and some who were not Hispanic/Latino may have been misclassified as such. After excluding patients who identified as Hispanic/Latino our results were unchanged, but further research is important to evaluate potential variations in care. In particular, future researchers should attempt to use the gold-standard of self-report when evaluating race/ethnicity for providers as well, as this is the most accurate way to capture this factor. Our study was limited by the data available in the EMR, so we are missing some variables, such as duration of symptoms or immunosuppression, which may be important. We limited our sample to visits where an attending physician had direct interaction with the patient. However, some attending physicians may have been accompaniefd by medical students, residents, or fellows. While our sample size was too small to examine these different potential pairings, future research should examine how the presence of trainees affects the prescription of antibiotics for acute bronchitis. Identification of bronchitis relied on ICD-9 codes in the medical record, which may be imperfect. A sample of chart reviews confirmed ICD-
74 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for 9 coding accuracy and past research suggests that these codes reliably identify bronchitis (Maselli, Gonzales, and Colorado Medical Society Joint Data Project 2001). CONCLUSION In conclusion, patient-provider race concordance was associated with reduced antibiotic prescribing. While gender concordance was not significant, the direction was the same and it may prove significant in a larger study. Race concordance is not the only aspect of the patientprovider relationship, but it may measure elements important to that relationship such as cultural competency, patient-centered explanations of care, or trust. The findings can be used to inform efforts to improve antimicrobial stewardship interventions, which to date have had limited success and have largely focused on provider knowledge and behavior (Davey et al. 2013). Interventions directed towards the patient have been less frequent and often are limited to provision of informational materials (Davey et al. 2013). Future research should explore the mechanisms behind the effect of patient-provider race concordance, perhaps through qualitative interviews of patient/provider dyads. Non-clinical indicators such as race and sex concordance may enhance our understanding of how to combat unnecessary prescribing and inform the design of future interventions. FUNDING SOURCE This project was supported in part by the Agency for Healthcare Research and Quality (Grant #T32HS022242) and by a dissertation support grant from the Jayne Koskinas Ted Giovanis Foundation for Health and Policy, a private foundation based in Highland, Maryland, dedicated to effecting change in the health care industry for the greater public good. ACKNOWLEDGEMENTS The opinions, findings, and conclusions or recommendations expressed in this material are those of the authors, and not necessarily those of the Jayne Koskinas Ted Giovanis Foundation for Health and Policy, its directors, officers, or staff. The funding source had no role in the design, conduct, collection, management, analysis, or interpretation of the data or in the preparation, review, or approval of the manuscript. REFERENCES Barlam, T. F., J. R. Morgan, L. M. Wetzler, C. L. Christiansen, and M. L. Drainoni. (2015). Antibiotics for respiratory tract infections: a comparison of prescribing in an outpatient setting. Infect Control Hosp Epidemiol 36(2): 153-9. Barlam, T. F., Soria-Saucedo, R., Cabral, H. J., & Kazis, L. E. (2016). Unnecessary Antibiotics for Acute Respiratory Tract Infections: Association With Care Setting and Patient Demographics. Open Forum Infect Dis, 3(1). Bertakis, K. D. and R. Azari. (2012). Patient-centered care: the influence of patient and resident physician gender and gender concordance in primary care. Journal of Women's Health 21(3): 326-33. Centers for Disease Control and Prevention. 2014. "Antibiotic Resistance Threats in the United States, 2013. Available at: http://www.cdc.gov/drugresistance/threat-report-2013/. Accessed 28 March 2017
75 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Charlson, M. E., P. Pompei, K. L. Ales, and C. R. MacKenzie. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Diseases 40(5): 373-83. Coenen, S., Francis, N., Kelly, M., Hood, K., Nuttall, J., Little, P., Verheij, T. J., Melbye, H., Goossens, H., Butler, C. C. (2013). Are Patient Views about Antibiotics Related to Clinician Perceptions, Management and Outcome? A Multi-Country Study in Outpatients with Acute Cough. PLoS One 8(10) e76691. Cooper-Patrick, L., J. J. Gallo, J. J. Gonzales, H. T. Vu, N. R. Powe, C. Nelson, and D. E. Ford. (1999). Race, gender, and partnership in the patient-physician relationship. Journal of the American Medical Association 282(6): 583-9. Cummings, P. 2009. Methods for estimating adjusted risk ratios. Stata Journal 9: 175-96. Davey, P., E. Brown, E. Charani, L. Fenelon, I. M. Gould, A. Holmes, C. R. Ramsay, P. J. Wiffen, and M. Wilcox. (2013). Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev 4: CD003543. Elliott, M. N., A. Fremont, P. A. Morrison, P. Pantoja, and N. Lurie. (2008). A new method for estimating race/ethnicity and associated disparities where administrative records lack self-reported race/ethnicity. Health Services Research 43(5 Pt 1): 1722-36. Ellis, H. D. and J. B. Deregowski. (1981). Within-race and between-race recognition of transformed and untransformed faces. The American journal of psychology 94(1): 27-35. Fiscella, K. and D. R. Williams. (2004). Health disparities based on socioeconomic inequities: implications for urban health care. Academic medicine : journal of the Association of American Medical Colleges 79(12): 1139-47. Grant, R. W., J. M. Ashburner, C. S. Hong, Y. Chang, M. J. Barry, and S. J. Atlas. (2011). Defining patient complexity from the primary care physician's perspective: a cohort study. Ann Intern Med 155(12): 797-804. Hicks, L. A., Bartoces, M. G., Roberts, R. M., Suda, K. J., Hunkler, R. J., Taylor, T. H., Jr., & Schrag, S. J. (2015). US outpatient antibiotic prescribing variation according to geography, patient population, and provider specialty in 2011. Clin Infect Dis, 60(9): 1308-1316. Jerant, A., K. D. Bertakis, J. J. Fenton, D. J. Tancredi, and P. Franks. (2011). Patient-provider sex and race/ethnicity concordance: a national study of healthcare and outcomes. Med Care 49(11): 1012-20. Kroening-Roche, J. C., A. Soroudi, E. M. Castillo, and G. M. Vilke. (2012). Antibiotic and bronchodilator prescribing for acute bronchitis in the emergency department. J Emerg Med 43(2): 221-7. Mangione-Smith, R., Stivers, T., Elliott, M., McDonald, L., & Heritage, J. 2003. Online commentary during the physical examination: a communication tool for avoiding inappropriate antibiotic prescribing? Social Science & Medicine, 56(2): 313-320. Maselli, J. H., R. Gonzales, and I. Colorado Medical Society Joint Data Project. (2001). Measuring antibiotic prescribing practices among ambulatory physicians: accuracy of administrative claims data. Journal of Clinical Epidemiology 54(2): 196-201. Shapiro, D. J., Hicks, L. A., Pavia, A. T., & Hersh, A. L. (2013). Antibiotic prescribing for adults in ambulatory care in the USA, 2007-09. Journal of Antimicrobial Chemotherapy 69(1): 234-240.
76 Patient-Provider Race and Sex Concordance: New Insights into Antibiotic Prescribing for Sohn, M. W., H. Zhang, N. Arnold, K. Stroupe, B. C. Taylor, T. J. Wilt, and D. M. Hynes. (2006). Transition to the new race/ethnicity data collection standards in the Department of Veterans Affairs. Population Health Metrics 4: 7. Steinman, M. A., C. S. Landefeld, and R. Gonzales. (2003). Predictors of broad-spectrum antibiotic prescribing for acute respiratory tract infections in adult primary care. Journal of the American Medical Association 289(6): 719-25. Tell, D., Engström, S., & Mölstad, S. (2015). Adherence to guidelines on antibiotic treatment for respiratory tract infections in various categories of physicians: a retrospective crosssectional study of data from electronic patient records. BMJ Open 5(7): e008096. Tonkin-Crine, S., Anthierens, S., Francis, N. A., Brugman, C., Fernandez-Vandellos, P., Krawczyk, J., Llor, C., Yardley, L., Coenen, S., Godycki-Cwirko, M., Butler. C. C., Verheij, T. J., Goossens, H., Little, P., Cals, J. W. (2014). Exploring patients views of primary care consultations with contrasting interventions for acute cough: a six-country European qualitative study. Primary Care Respiratory Medicine 24: 14026. Vaz, L. E., Kleinman, K. P., Lakoma, M. D., Dutta-Linn, M. M., Nahill, C., Hellinger, J., & Finkelstein, J. A. (2015). Prevalence of Parental Misconceptions About Antibiotic Use. Pediatrics 136(2), 221-231.